Thoughts on Render Farms and Boxx tech?
polycounter lvl 14
The company I work for is exploring options for a new render farm. Weve been using two custom built computers IT put together for us a while back as our farm and they just arent cutting it anymore. I love my IT friends and all, but they dont understand our needs. It takes almost two days to render 1,800 frames, roughly 2-6 minutes a frame depending on how close the camera is and other factors.
Weve been looking at a few of the RenderBoxx and RenderPro options, and we like what were seeing, but we cant really decide. First off, does anyone have any experience with using any of Boxxs products? Id love to hear about it, because its going to be my nuts if we invest in this and it doesnt perform as expected.
Were leaning towards RenderPro a little bit more because IT feels the need to not give us rights to our own farm. We cant even delete a job without calling them up. At least with the RenderPro we dont need IT around injecting their ideas of what is right and wrong, holy and unholy, into our daily lives. However, those little things dont seem to pack as much punch for the price as a single RenderBoxx does.
The way I see it, its one RenderBoxx (lowest option is 16 cores) or two RenderPros (8 cores each) for my coworker and I. Wed probably try and network our RenderPros together between our two cubes. Im honestly not sure which one at this point.
Anyway, enough blabbing. Who here uses a farm and what experience can you share with me? I suppose well stop using backburner too and opt for Qube! or something. Backburner doesnt seem very robust if the task size in Maya is larger than 25, the farms seem to time out and the whole task gets deleted (many frames), causing us to lose time on the project. Once one render node goes out, it seems like thats it. Theres no getting it back on board without restarting the job.
Weve been looking at a few of the RenderBoxx and RenderPro options, and we like what were seeing, but we cant really decide. First off, does anyone have any experience with using any of Boxxs products? Id love to hear about it, because its going to be my nuts if we invest in this and it doesnt perform as expected.
Were leaning towards RenderPro a little bit more because IT feels the need to not give us rights to our own farm. We cant even delete a job without calling them up. At least with the RenderPro we dont need IT around injecting their ideas of what is right and wrong, holy and unholy, into our daily lives. However, those little things dont seem to pack as much punch for the price as a single RenderBoxx does.
The way I see it, its one RenderBoxx (lowest option is 16 cores) or two RenderPros (8 cores each) for my coworker and I. Wed probably try and network our RenderPros together between our two cubes. Im honestly not sure which one at this point.
Anyway, enough blabbing. Who here uses a farm and what experience can you share with me? I suppose well stop using backburner too and opt for Qube! or something. Backburner doesnt seem very robust if the task size in Maya is larger than 25, the farms seem to time out and the whole task gets deleted (many frames), causing us to lose time on the project. Once one render node goes out, it seems like thats it. Theres no getting it back on board without restarting the job.